Regression analysis of instrumental texture characteristics and sensory characteristics of pork

In order to analyze the correlation between sensory attributes and instrumental texture characteristics of pork, and validate model for predict sensory attributes, two instrumental methods for assessing texture characteristics of pork, warner–bratzler (W-B) and texture profile analysis (TPA), were conducted on 54 samples of m. longissimus dorsi muscle from 18 DLY commercial pigs, aging at 24, 72 and 144 hours post-mortem. A trained panelists sensory analysis was also performed on the similar 54 samples. The results showed that, most of the instrumental measures were significantly or marked significantly correlated (P<0.05 or P<0.01) with sensory texture characteristics. Principal component-stepwise regression analysis were used to generate prediction equations with the data of TPA and warner-bratzler measures as independent variables respectively and the data of sensory analysis as dependent variable. The prediction equations of sensory hardness, springiness and juiciness were of significance in statistics and the determination coefficients were 61.46%, 39.77% and 42.34%, respectively. Therefore, the widespread correlation between sensory attributes and instrumental texture characteristics of pork are comfirmed, and the prediction efficiency of meat sensory attributes prediction model is brought to a new level by more proper testing and statistical methods.